Closed dhmemi closed 5 months ago
The ci failed, but where I can get the details? The actions' outputs seem unhelpful.
It is no longer a problem generated by your code. The problem is coming from the github-runner which are using cmake 3.27 and nlohmann::json which allows CMake 3.1. This is no longer compatibile.
I really don't know what to do now.
@pboettch I'd recommend trying to use pip3
to install a specific version of CMake in your CI script, rather than using the current template. Something like:
- name: Get cmake == 3.24.1
run: |
apt install -y python3-pip
pip3 install cmake==3.24.1
@pboettch I'd recommend trying to use
pip3
to install a specific version of CMake in your CI script, rather than using the current template. Something like:- name: Get cmake == 3.24.1 run: | apt install -y python3-pip pip3 install cmake==3.24.1
All of the CI workflow has only been integrated but done by others. I'm grateful, but have to admin, that I don't really know where to put it. Could you help out with a PR?
Thanks.
Well, maybe i found how to do it. See #296
All of the CI workflow has only been integrated but done by others. I'm grateful, but have to admin, that I don't really know where to put it. Could you help out with a PR?
I'll see what I can do to help out - I have a decent amount of experience with various CI systems
I'm closing this one in favor of #297
Could you please rebase your code on top of master and run tests locally. The multipleOf-tests fails for all versions for me locally.
I rebased and pushed to your branch to your repo.
fixed.
As the multipleOf_.second
is always float type, the method of calculating the error by rounding the quotient always works.
Seems still buggy, do not merge even if ci passed, I'll add more cases.
This branch is ready to be merged if there are no further suggested modifications.
I have adjusted the logic for determining whether a floating-point number is a multiple of another number based on the difference in multiples of the values specified in the value
and multipleOf
fields. While the criteria for such determination have become less strict, this should be more practical for most cases. It helps to avoid validation failures caused by slight floating-point drift resulting from simple numerical calculations.
Thanks a lot for your work.
Closes #293
Fix validating of float-point value.